---
title: "Data Science Transport – Second Assignment – Group 12"
output: 
  flexdashboard::flex_dashboard:
    orientation: columns
    vertical_layout: fill
    source_code: embed
---

```{r setup, include=FALSE}
library(flexdashboard)
library(gtfsrouter)
library(tidyverse)
library(tidytransit)
library(sf)
library(tmap)
library(units)
library(RColorBrewer)
tmap_mode("view")
```

last time {data-icon="fa-hourglass-half"}
=====================================
[Last time](https://bernd886.github.io/data-science-transport-mid-assign-wise-2021/) we took a look at the GTFS data of the VBB and the output data of car-free zone MATSim simulation.
We stuck to the data sets. This time you will see plots showing results of :

- GTFS: routing on the whole pt network (last time only BVG and S-Bahn)
- MATSim: emissions distribution

TO EDIT

ADD DATA SOURCES!!!!!!!!!!!!!!!

- test1
- test2

### Chart 0
```{r}

library(leaflet)
leaflet() %>%
  addTiles() %>%
  addMarkers(lng=174.768, lat=-36.852, popup="The birthplace of R")
```

### Chart 1

```{r}
# 1. Plot the dots themselves
```

### Cars

```{r}
knitr::kable(mtcars)
```

isochrones {data-icon="fa-expand-arrows-alt"}
=====================================

```{r, include = FALSE}
##############################################################
#
#   READ GTFS DATA
#
##############################################################
# set work directions
setwd_gtfs <- function(){setwd("~/Documents/Uni/Master/DataScienceTransport/data/vbb-gtfs")}
setwd_data <- function(){setwd("~/Documents/Uni/Master/DataScienceTransport/data")}
setwd_work <- function(){setwd("~/Documents/Uni/Master/DataScienceTransport/assignment_2")}

setwd_work

# read gtfs data for monday
file <- file.path("~/Documents/Uni/Master/DataScienceTransport/data/vbb-gtfs/2020-12_2020-12-28.zip")
gtfs <- extract_gtfs(file) %>% gtfs_timetable(day = 2)

##############################################################
#
#   SET TIMES
#
##############################################################
start_time <- 7 * 3600 + 1200
end_time <- 8 * 3600

# create isochrone
# ic <- gtfs_isochrone (gtfs,
#                       from = from,
#                       start_time = start_time,
#                       end_time = end_time)

##############################################################
#
#   CREATE STOPS SF OBJECT
#
##############################################################
stops <- st_as_sf(gtfs$stops,
                   coords = c("stop_lon", "stop_lat"),
                   crs = 4326) %>% 
  st_transform(25833)

##############################################################
#
#   SHAPE DISTRICTS NEW (+ area)
#
##############################################################
setwd_data()
shape_districts_new <- read_sf(dsn = "LOR_SHP_2019-1", layer = "Planungsraum_EPSG_25833")
setwd_work()

shape_districts_new <- shape_districts_new %>% 
  group_by(BEZIRK) %>% 
  summarise() %>% 
  filter(!is.na(BEZIRK)) %>% 
  rename(NAME = BEZIRK) %>% 
  mutate(AREA = st_area(geometry)) %>% 
  select(NAME, AREA, everything()) %>% 
  mutate(AREA = (AREA / 1000000) * as_units("km2"))

# setting crs of polygons
st_crs(shape_districts_new$geometry) <- 25833

shape_berlin <- st_union(shape_districts_new)

##############################################################
#
#   SPECIFIC SHAPES AND STOPS
#
##############################################################

stops_in_berlin <- stops %>% 
  mutate(inside_berlin = st_within( geometry, shape_berlin )) %>% 
  mutate(inside_berlin = !is.na( as.numeric( inside_berlin ))) %>% 
  filter(inside_berlin == TRUE)

# get isochrone area
# ic = gtfs_isochrone (gtfs,
#                      from = "Berlin, Sowjetisches Ehrenmal",
#                      #from_is_id = TRUE,
#                      start_time = start_time,
#                      end_time = end_time)$hull$area

##############################################################
#
#   CALCULATE ISOCHRONES
#
##############################################################
# # the following code calculates the isochrones (inefficent, ~ 15h)
# # instead of running the code, we read in the pre-calculated file

# stops_ic_area <- vector(mode = "double")
# 
# # create isochrone areas for stops in 50 minutes
# for (stop_name in stops$stop_name){
# 
#   tryCatch( {
#     ic_area <- gtfs_isochrone (gtfs,
#                                from = stop_name,
#                                #from_is_id = TRUE,
#                                start_time = start_time,
#                                end_time = end_time)$hull$area
#     if(is.null(ic_area)) {
#       stops_ic_area <<- rbind(stops_ic_area, 0)
#       print(paste(stop_name, ": ", ic_area, "!!!!!!!!!!"))
#     } else {
#       stops_ic_area <<- rbind(stops_ic_area, ic_area)
#       print(paste(stop_name, ": ", ic_area))
#     }
#     },
#     error = function(e) {
#       stops_ic_area <<- rbind(stops_ic_area, 0)
#       print(paste("ERROR!!!", stop_name))
#       }
#     )
# }
# 
# ##############################################################
# #
# #   CLEANING
# #
# ##############################################################
# 
# # merge and clean
# # https://r-spatial.github.io/sf/reference/bind.html
# # https://cran.r-project.org/web/packages/units/vignettes/units.html
# rownames(stops_ic_area) <- NULL
# stops_area <- st_sf(data.frame(stops, stops_ic_area / 1000000)) %>%
#   rename(ic_area = stops_ic_area.1e.06,
#          id = stop_id,
#          name = stop_name,
#          parent = parent_station) %>% 
#   select(id, name, parent, ic_area) %>% 
#   mutate(ic_area = ic_area * as_units("km2"))
# 
# # save
# # https://r-spatial.github.io/sf/reference/st_write.html
# st_write(stops_area, "output_stops_ic_area.shp")
stops_area <- st_read("output_stops_ic_area.shp")

# more cleaning for plot
# https://dplyr.tidyverse.org/reference/distinct.html
stops_area = 
  stops_area %>% 
  select(name, ic_area) %>% 
  distinct(name, .keep_all = TRUE)

stops_area_berlin <- stops_area %>% 
  mutate(inside_berlin = st_within( geometry, shape_berlin )) %>% 
  mutate(inside_berlin = !is.na( as.numeric( inside_berlin ))) %>% 
  filter(inside_berlin == TRUE) %>% 
  select(-inside_berlin) %>% 
  mutate(id = paste(name, ": ", round(ic_area)))
```

Column {data-width=100}
-------------------------------------
    
### Approach {data-height=200}

The [gtfsrouter](https://atfutures.github.io/gtfs-router/) package allows us to calculate all stations reachable within a specified time period from a nominated station ([isochrones](https://atfutures.github.io/gtfs-router/reference/gtfs_isochrone.html)). We use the hull polygon as an indicator for the city area reachable.

See below for a isochrone plot of a 30min Monday work–home trip starting from Helmholtzstraße at 18:00.

We try to simulate a home–work trip in the morning rush hour arriving at 08:00. Assuming a five minute walk at beginning and end of the trip, we limit the GTFS data to Monday (2021-01-18) and calculate the isochrones for a 40min time range. By that we want to get all reachable stations by local transport (< 60min, [PBefG §8 (1)](https://www.gesetze-im-internet.de/pbefg/__8.html)) and rate the station inside the network.

- reachablity of possible work places
- calculation with all transport agencies
-- domination of the hull area in 40 min time range by reachable fanning out long-distance transport
- no weighting of the reachable area

### 30min IC home trip (18:00) from Helmholtzstr. {data-height=100}
    
```{r}
ic_einstein <- gtfs_isochrone(gtfs,
                              from = "Berlin, Helmholtzstr.",
                              start_time = 18 * 3600,
                              end_time = 18 * 3600 + 1800)

tm_basemap(leaflet::providers$OpenStreetMap.DE) +
  tm_shape(ic_einstein$hull) + 
  tm_polygons(col = "red",
              alpha = 0.2,
              border.col = "red") +
  tm_shape(ic_einstein$routes) +
  tm_lines() +
  tm_shape(ic_einstein$end_points) +
  tm_dots(col = "red") + 
  tm_shape(ic_einstein$start_point) + 
  tm_dots(col = "green")
```

Column {data-width=300}
-------------------------------------
   
### area size of the hull enclosing the routed points

```{r}
##############################################################
#
#   PLOT
#
##############################################################

tm_shape(shape_districts_new) +
  tm_polygons(alpha = 0,
              popup.vars = c("area" = "AREA")) +
  tm_shape(stops_area_berlin) +
  tm_dots(col = "ic_area",
          id = "name",
          popup.vars = c("area" = "ic_area"),
          size = 0.07,
          border.lwd = 0.3,
          legend.hist = TRUE,
          n = 15,
          title = "isochrone area [km^2]") +
  tm_view(bbox = shape_berlin)
```


traveltimes to center {data-icon="fa-stopwatch"}
=====================================

```{r, include = FALSE}
##############################################################
#
#   SHAPE CENTER AREAS
#
##############################################################
# "Zentrentragender Stadtraum mit höchster / hoher Urbanität"
# of Zentrumsbereichskernen
# see page 39: https://www.stadtentwicklung.berlin.de/planen/stadtentwicklungsplanung/download/zentren/2011-07-31_StEP_Zentren3.pdf
# or page 45 (less detailed): https://www.stadtentwicklung.berlin.de/planen/stadtentwicklungsplanung/download/zentren/StEP_Zentren_2030.pdf
# recreated with QGis

shape_center <- read_sf(dsn = "shape_center_areas", layer = "center_areas") %>% 
  mutate(name = c("east", "west")) %>% 
  select(name)

shape_center_east <- shape_center %>% filter(name == "east")
shape_center_west <- shape_center %>% filter(name == "west")

##############################################################
#
#   READ GTFS DATA
#
##############################################################
# now we work with tidytransit
# calculation of shortest tt from all station to specific ones is more convinent

setwd_gtfs()
gtfs <- read_gtfs("2020-12_2020-12-28.zip")
setwd_work()

# http://tidytransit.r-transit.org/reference/filter_stop_times.html
stop_times_filtered <- filter_stop_times(gtfs, "2021-01-18", "06:00:00", "07:55:00")

##############################################################
#
#   GET STOPS
#
##############################################################
stops <- st_as_sf(gtfs$stops, coords = c("stop_lon", "stop_lat"), crs = 4326) %>%
  st_transform(25833) %>% 
  select(stop_name) %>%
  rename(name = stop_name) %>%
  distinct(name)

stops_berlin <- stops %>% 
  mutate(inside_berlin = st_within( geometry, shape_berlin )) %>% 
  mutate(inside_berlin = !is.na( as.numeric( inside_berlin ))) %>% 
  filter(inside_berlin == TRUE) %>% 
  select(name)

stops_center <- stops %>% 
  mutate(inside_center = st_within( geometry, shape_center )) %>% 
  mutate(inside_center = !is.na( as.numeric( inside_center ))) %>% 
  filter(inside_center == TRUE) %>% 
  select(name)

stops_center_east <- stops %>% 
  mutate(inside_center_east = st_within( geometry, shape_center_east )) %>% 
  mutate(inside_center_east = !is.na( as.numeric( inside_center_east ))) %>% 
  filter(inside_center_east == TRUE) %>% 
  select(name)

stops_center_west <- stops %>% 
  mutate(inside_center_west = st_within( geometry, shape_center_west )) %>% 
  mutate(inside_center_west = !is.na( as.numeric( inside_center_west ))) %>% 
  filter(inside_center_west == TRUE) %>% 
  select(name)

##############################################################
#
#   TT calculation
#
##############################################################
# what are the tt to the center areas?
# according to Nahverkehrsplan Berlin 2019-2023: ANlage 1 - Monitoringbericht (p. 12)
# standard: tt_max = 3600, n_transfer_max = 2, n_realise_stations = 0.95

tt <- travel_times(
  stop_times_filtered,
  stops_center$name,
  time_range = 5400,
  arrival = TRUE,
  max_transfers = 2,
  # max_departure_time = NULL,
  return_coords = TRUE,
  return_DT = FALSE
)

# clean it for plot
tt <- tt %>% 
  rename(from = from_stop_name,
         to = to_stop_name,
         tt = travel_time,
         departure = journey_departure_time,
         arrival = journey_arrival_time
         ) %>% 
  select(-c(from_stop_id, to_stop_id, to_stop_lat, to_stop_lon)) %>% 
  st_as_sf(coords = c("from_stop_lon", "from_stop_lat"),
           crs = 4326) %>% 
  st_transform(25833) %>% 
  mutate(tt = set_units(round(tt/60, 2), "min"))
```

Column {data-width=100}
-------------------------------------
    
### approach {data-height=400}

The [local transport plan](https://www.berlin.de/sen/uvk/verkehr/verkehrsplanung/oeffentlicher-personennahverkehr/nahverkehrsplan/) (p. 106) sets targets for the connectivity standards. Different categories of center areas (see [StEP](https://stadtentwicklung.berlin.de/planen/stadtentwicklungsplanung/de/zentren/zentren2030/index.shtml), p. 45) have to be reachable within a certain time and with a maximum number of transfer. This should hold for 95% of the stations.

Based on the GTFS data, we try to recreate the result of the monitoring ([NVP Anlage 1](https://www.berlin.de/sen/uvk/_assets/verkehr/verkehrsplanung/oeffentlicher-personennahverkehr/nahverkehrsplan/broschure_nvp_2019_anlage_1.pdf), p. 12), mentioning a degree of fulfillment of 99,7% for the central areas.

* destination:
  + City West (Zoo/ Kurfürstendamm)
  + Mitte (Potsdamer Platz/ Alexanderplatz)
* max. tt: 60min
* max. transfers: 2

From a more detailed [display of the area](https://www.stadtentwicklung.berlin.de/planen/stadtentwicklungsplanung/download/zentren/2011-07-31_StEP_Zentren3.pdf) (p. 39), we create a shape file enclosing the associated stations. The [tidytransit](https://tidytransit.r-transit.org/) package let's us [calculate](https://tidytransit.r-transit.org/reference/travel_times.html) the shortest travel time for all stations to any of a specified set of stations. For that `arrival` has to be set to `TRUE`.

We try to simulate a home–work trip in the morning rush hour arriving at 08:00. Assuming five minute walking at beginning and end of the trip, we limit the GTFS data to Monday (2021-01-18) and calculate the shortest travel times in a 90min time range ending at 07:55.

### Percent of all berlin stations fullfill the connectivity standard according to the GTFS data {data-height=200}

```{r}
##############################################################
#
#   DEGREE OF FULLFILMENT
#
##############################################################

n_of_stations <- tt %>%
  mutate(inside_berlin = st_within( geometry, shape_berlin )) %>% 
  mutate(inside_berlin = !is.na( as.numeric( inside_berlin ))) %>% 
  filter(inside_berlin == TRUE) %>% 
  mutate(outside_center = st_within( geometry, shape_center )) %>% 
  mutate(outside_center = is.na( as.numeric( outside_center ))) %>% 
  filter(outside_center == TRUE) %>%
  nrow()

n_of_stations_valid <- tt %>% 
  mutate(inside_berlin = st_within( geometry, shape_berlin )) %>% 
  mutate(inside_berlin = !is.na( as.numeric( inside_berlin ))) %>% 
  filter(inside_berlin == TRUE) %>% 
  mutate(outside_center = st_within( geometry, shape_center )) %>% 
  mutate(outside_center = is.na( as.numeric( outside_center ))) %>% 
  filter(outside_center == TRUE) %>%
  filter(tt <= 60 * as_units("min")) %>% 
  filter(transfers <= 2) %>% 
  nrow()

percent_stations_valid <- n_of_stations_valid / n_of_stations * 100
percent_stations_valid <- round(percent_stations_valid, 2)

valueBox(paste(percent_stations_valid, "%"), icon = "fa-crosshairs")
```
    

Column {data-width=300}
-------------------------------------
   
### shortest travel time to one of the stations inside City West or Mitte

```{r}
##############################################################
#
#   PLOT
#
##############################################################

# https://campus.datacamp.com/courses/visualizing-geospatial-data-in-r/raster-data-and-color?ex=9
rdylgn <- rev(brewer.pal(7, "RdYlGn"))

# https://leaflet-extras.github.io/leaflet-providers/preview/
# https://tlorusso.github.io/geodata_workshop/tmap_package
# https://www.rdocumentation.org/packages/tmap/versions/3.0/topics/tm_basemap
# https://rdrr.io/cran/tmap/man/tm_view.html
# https://leafletjs.com/reference-1.3.4.html#map-methods-for-modifying-map-state

tm_basemap(leaflet::providers$CartoDB.DarkMatter) +
  tm_shape(shape_districts_new) + 
  tm_polygons(alpha = 0,
              lwd = 1.5,
              border.col = "white",
              popup.vars = c("area" = "AREA")
              ) +
  tm_shape(shape_center) +
  tm_polygons(alpha = 0.2,
              col = "red",
              border.col = "red"
              ) + 
  tm_shape(tt) +
  tm_dots(col = "tt",
          style = "fixed",
          breaks = c(0, 10, 20, 30, 40, 50, 60, 120),
          labels = c("0 – 10", "10 – 20", "20 – 30", "30 – 40", "40 – 50", "50 – 60", "> 60"), 
          id = "from",
          palette = rdylgn,
          title = "traveltime [min]",
          popup.vars = c("to" = "to", 
                         "traveltime" = "tt",
                         "departure at" = "departure",
                         "arrival at" = "arrival",
                         "number of transfers" = "transfers")
          ) +
  tm_view(bbox = shape_center)
```

Hao 1 {data-icon="fa-burn"}
=====================================

Column {data-width=100}
-------------------------------------

### PUT TEXT HERE OR NOT

Lorem ipsum dolor sit amet, consectetur adipiscing elit. In vel turpis lacus. Vivamus enim magna, euismod sit amet mi nec, malesuada euismod nulla. Phasellus non sem euismod, porta neque sit amet, aliquam lorem. Aenean at gravida magna. Morbi magna lorem, vestibulum in elit a, maximus lacinia massa.

Phasellus non tincidunt erat, sit amet volutpat sem. Sed maximus purus ut libero rutrum fringilla. Nullam lobortis ullamcorper purus lacinia eleifend.

In dictum enim urna, at ultricies lectus luctus id. Aenean vulputate lectus urna, pellentesque ultrices libero feugiat eu.

Column {data-width=300}
-------------------------------------

```{r, include = FALSE}
berlin_bezirke <- st_read("~/Documents/Uni/Master/DataScienceTransport/assignment_2/Hao/shp-bezirke/bezirke_berlin.shp")

berlin_emissions <-read_delim("~/Documents/Uni/Master/DataScienceTransport/assignment_2/Hao/berlin-v5.5-1pct.emissionsgrid_Berlin_PlanA.csv", 
                            delim="\t",
                            # sep = "\t",
                            locale=locale(decimal_mark = "."),
                            col_types = cols(
                              x = col_double(),
                              y = col_double()
                            ))
```

### PUT TEXT HERE OR NOT

```{r}
map = read_sf("~/Documents/Uni/Master/DataScienceTransport/assignment_2/Hao/shp-bezirke/bezirke_berlin.shp")

berlin_emissions_sf <- st_as_sf(berlin_emissions, coords = c('x', 'y'), crs = st_crs(map))

berlin_emissions_bezirke <- berlin_emissions_sf %>% mutate(
  intersection = as.integer(st_intersects(geometry, map))
  , area = if_else(is.na(intersection), '', map$Name[intersection])
) 

# ----
joined <- st_join(berlin_bezirke, berlin_emissions_bezirke)

# Plot1
joined_count <- joined %>%
  group_by(Name) %>%
  summarise(sum=sum(PM2_5_non_exhaust)) %>%
  #summarise(sum=sum(N2O)) %>%
  #summarise(sum=sum(NH3)) %>%
  ungroup() %>%
  mutate(col_value=sum/n()) %>% 
  #mutate(col_value=1/sum*n())
  rename("Particular Matter" = "col_value")
  
tm_shape(joined_count) +
  # tm_borders() +
  tm_polygons(col="Name") +
  tm_shape(joined_count) +
  # tm_dots(size="size_value", col="col_value")
  tm_symbols(size="sum", col="Particular Matter")
```


Hao 2 {data-icon="fa-car"}
=====================================

Column {data-width=100}
-------------------------------------

### PUT TEXT HERE OR NOT

Lorem ipsum dolor sit amet, consectetur adipiscing elit. In vel turpis lacus. Vivamus enim magna, euismod sit amet mi nec, malesuada euismod nulla. Phasellus non sem euismod, porta neque sit amet, aliquam lorem. Aenean at gravida magna. Morbi magna lorem, vestibulum in elit a, maximus lacinia massa.

Phasellus non tincidunt erat, sit amet volutpat sem. Sed maximus purus ut libero rutrum fringilla. Nullam lobortis ullamcorper purus lacinia eleifend.

In dictum enim urna, at ultricies lectus luctus id. Aenean vulputate lectus urna, pellentesque ultrices libero feugiat eu.

Column {data-width=300}
-------------------------------------

### PUT TEXT HERE OR NOT

```{r}
# Plot2
joined_count <- joined %>%
  group_by(Name) %>%
  #summarise(sum=sum(PM2_5_non_exhaust)) %>%
  summarise(sum=sum(N2O)) %>%
  #summarise(sum=sum(NH3)) %>%
  ungroup() %>%
  #mutate(col_value=sum/n())
  mutate(col_value=1/sum*n())
# view(joined_count)

berlin_emissions_bezirke <- berlin_emissions_bezirke %>% 
  rename("Nitrous Oxide (N2O)" = "N2O")
  
tmap_mode("plot")
tm_shape(joined_count) +
  # tm_borders() +
  tm_polygons() +
  # tm_polygons(col="col_value") +
  # tm_dots(size="size_value", col="col_value")
  tm_shape(berlin_emissions_bezirke) +
  tm_dots(size=0.01, col="Nitrous Oxide (N2O)", border.lwd=NA)
  # tm_dots(size=0.001, col="N2O", alpha=0.1)
```

sources {data-icon="fa-external-link-alt"}
=====================================
***


* VBB GTFS data:
  + [Website](https://www.vbb.de/unsere-themen/vbbdigital/api-entwicklerinfos/datensaetze)
  + [use data set](http://transitfeeds.com/p/verkehrsverbund-berlin-brandenburg/213/20201228) on transitfeeds
* Berlin shape files: [Senatsverwaltung für Stadtentwicklung und Wohnen](https://web.archive.org/web/20190624123508/https://www.stadtentwicklung.berlin.de/planen/basisdaten_stadtentwicklung/lor/de/download.shtml)
* GTFS routing:
  + [tidytransit](https://tidytransit.r-transit.org/) package
  + [gtfs-router](https://atfutures.github.io/gtfs-router/) package
* helpful references:
  + sf reference: [bind](https://r-spatial.github.io/sf/reference/bind.html), [st_write](https://r-spatial.github.io/sf/reference/st_write.html), [st_read](https://r-spatial.github.io/sf/reference/st_read.html), 
  + [Units of Measurement for R Vectors](https://cran.r-project.org/web/packages/units/vignettes/units.html)
  + tidyverse dplyr: [distinct](https://dplyr.tidyverse.org/reference/distinct.html)
* LITERATUR?!
  + [Stadtentwicklungsplan Zentren 3](https://www.stadtentwicklung.berlin.de/planen/stadtentwicklungsplanung/download/zentren/2011-07-31_StEP_Zentren3.pdf) (2011)
  + [Stadtentwicklungsplan Zentren 2030](https://www.stadtentwicklung.berlin.de/planen/stadtentwicklungsplanung/download/zentren/StEP_Zentren_2030.pdf) (2019)
  + [Nahverkehrsplan 2019-2023](https://www.berlin.de/sen/uvk/verkehr/verkehrsplanung/oeffentlicher-personennahverkehr/nahverkehrsplan/) (2019)
* plotting:
  + datacamp: [Custom palette in tmap](https://campus.datacamp.com/courses/visualizing-geospatial-data-in-r/raster-data-and-color?ex=9)
  + [Leaflet-providers preview](https://leaflet-extras.github.io/leaflet-providers/preview/)
  + tlorusso: [tmap](https://tlorusso.github.io/geodata_workshop/tmap_package)
  + rdocumentation: [tm_basemap](https://www.rdocumentation.org/packages/tmap/versions/3.0/topics/tm_basemap)
  + tm_view: [Options for the interactive tmap viewer](https://rdrr.io/cran/tmap/man/tm_view.html)